Short-term reforecasting of power output from a 48 MWe solar PV plant

•Smart reforecast is applied to the intra-hour power prediction of a PV plant.•Predictions of generation from three baseline models are analyzed and reforecasted.•The reforecast models are based on ANN optimized by genetic algorithm.•The reforecasts significantly improve the forecast skills of the b...

Full description

Saved in:
Bibliographic Details
Published in:Solar energy Vol. 112; pp. 68 - 77
Main Authors: Chu, Yinghao, Urquhart, Bryan, Gohari, Seyyed M.I., Pedro, Hugo T.C., Kleissl, Jan, Coimbra, Carlos F.M.
Format: Journal Article
Language:English
Published: New York Elsevier Ltd 01.02.2015
Pergamon Press Inc
Subjects:
Artificial neural networks
Clouds
Forecasting
Genetic algorithm optimization
Neural networks
Photovoltaic cells
Photovoltaics
Position tracking
Prediction models
PV generation
Real-time reforecasting
Solar energy
Tracking techniques
Real-time reforecasting
PV generation
Genetic algorithm optimization
Artificial neural networks
ISSN:0038-092X, 1471-1257
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Smart reforecast is applied to the intra-hour power prediction of a PV plant.•Predictions of generation from three baseline models are analyzed and reforecasted.•The reforecast models are based on ANN optimized by genetic algorithm.•The reforecasts significantly improve the forecast skills of the baseline models. A smart, real-time reforecast method is applied to the intra-hour prediction of power generated by a 48 MWe photovoltaic (PV) plant. This reforecasting method is developed based on artificial neural network (ANN) optimization schemes and is employed to improve the performance of three baseline prediction models: (1) a physical deterministic model based on cloud tracking techniques; (2) an auto-regressive moving average (ARMA) model; and (3) a k-th Nearest Neighbor (kNN) model. Using the measured power data from the PV plant, the performance of all forecasts is assessed in terms of common error statistics (mean bias, mean absolute error and root mean square error) and forecast skill over the reference persistence model. With the reforecasting method, the forecast skills of the three baseline models are significantly increased for time horizons of 5, 10, and 15min. This study demonstrates the effectiveness of the optimized reforecasting method in reducing learnable errors produced by a diverse set of forecast methodologies.
Bibliography:SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ISSN:0038-092X
1471-1257
DOI:10.1016/j.solener.2014.11.017